African heritage sites threatened as sea-level rise accelerates

Heritage sites provide substantial cultural, ecological, historical, social and economic value but are increasingly threatened by climate hazards. While sea levels have accelerated over recent decades and are projected to continue rising through the 21st century, intensifying coastal flooding and erosion, there is a notable knowledge gap for Africa’s approximately 300,000 km of coastline across 39 countries. Few continent-wide assessments exist for the exposure of African cultural and natural heritage to coastal hazards. This study addresses that gap by quantifying current and future exposure of African heritage sites (AHS) to 1-in-100-year coastal flooding and erosion events under moderate (RCP 4.5) and high (RCP 8.5) emission scenarios, establishing a geospatial database of coastal AHS and projecting how exposure evolves this century.

Prior work has shown global and regional threats of climate change to heritage, including risks to UNESCO sites from coastal flooding and erosion and potential losses of inhabited places due to sea-level rise. Studies have reported accelerating global sea-level rise and intensifying coastal extremes driven by changing tides, surges and wave climate, as well as shoreline retreat along sandy coasts. However, Africa is under-represented in global heritage risk analyses and has relatively sparse assessments compared to other continents. Existing literature also highlights the role of natural coastal protections (coral reefs, seagrasses, mangroves) and their vulnerability to warming, acidification and anthropogenic pressures, which may exacerbate exposure in Africa. This work builds on and extends these strands by providing the first continent-wide, digitized, geospatial assessment of African coastal heritage exposure to combined coastal flooding and erosion under future climate scenarios.

• Heritage sites database: Compiled a continent-wide geospatial database of 284 coastal African heritage sites (213 natural, 71 cultural), drawing from UNESCO World Heritage List (including Tentative List) and the Ramsar Sites Information Service. Where polygons/maps were missing or imprecise (especially for World Heritage Sites and Tentative List entries), boundaries were delineated using the World Database on Protected Areas, Google Earth overlays, historical imagery and relevant literature. Site attributes include name, official designation number, centroid, polygon, area, and elevation statistics. Sites below 50 m elevation (GLO-90 DEM) were included to conservatively capture potential sea-level exposure.
• Scenarios and time horizon: Exposure evaluated for baseline (2010) and through the 21st century under RCP 4.5 and RCP 8.5 at decadal steps, with results emphasized for 2050 and 2100. Uncertainty represented via percentiles (5th–95th very likely range), reporting median values.
• Extreme sea levels (ESLs): Waves and storm surges hindcast (1980–2015) with DFLOW FM (surge) and WW3 (waves) driven by ECMWF ERA-Interim; tropical cyclones represented using IBTRACS best tracks and satellite altimetry for significant wave heights. Present-day tides from FES2014; probabilistic sea-level rise projections used to adjust tidal elevations. Future waves/surges forced by six CMIP5 climate models.
• Coastal inundation modelling: Inundation simulated using LISFLOOD-FP over coastal segments spaced every 25 km and extending up to 200 km inland, using GLO-90 DEM and land-use-derived hydraulic roughness. ESLs combine mean sea level, astronomical tide and meteorological tide (storm surge + wave setup) via Monte Carlo sampling and non-stationary extreme value analysis to derive median 100-year water levels. Areas permanently inundated by present-day high tide were excluded from exposure. No explicit consideration of coastal protection standards was made, assuming none reach the 100-year protection threshold at studied sites.
• Coastal erosion projections: Probabilistic shoreline change projections (to 2100, 10-year steps) combine (1) ambient shoreline dynamics, (2) retreat from relative sea-level rise (modified Bruun rule), and (3) episodic storm-driven erosion (100-year events). Post-processing constrained retreat where non-erodible lithology or coastal defences/obstructions exist, using the Global Lithological Map and inspection of Google Earth time-series imagery. This yielded 6 cultural and 55 natural sites potentially exposed to erosion.
• Combined hazard and metrics: For each site, scenario, year and percentile, exposed area from flooding and erosion were computed and the total exposed area was taken as the union/maximum of the two hazards. Exposure classes were defined by percentage of site area affected: none, small (<25%), moderate (25–50%), high (50–75%), very high (>75%). Results were aggregated to country, regional and continental levels.

• Baseline exposure (2010): 56 of 284 sites (20%) are exposed to a 1-in-100-year coastal extreme event: 35/213 natural (16%) and 21/71 cultural (30%). Total exposed area: 1,719 km² (natural 1,300 km²; cultural 419 km²). Average exposure across sites: 4.5%. Fifty sites have <50% of their area exposed; 3 sites >50%; 3 sites >75%.
• Regional distribution (current): North Africa has 23 exposed sites (of 109), West Africa 18, Southern Africa 7, East Africa 4, SIDS 4; no Central African sites currently exposed.
• 2050 projections: Number of exposed sites more than triples under moderate emissions to 191 (very likely range 191–196) and increases to 198 (198–210) under high emissions. Exposed area: RCP 4.5: 1,744 km² (1,728–3,681); RCP 8.5: 2,171 km² (1,757–3,932). Relative to baseline, additional exposed area is <2% (RCP 4.5) and ~25% (RCP 8.5).
• 2100 projections: The number of exposed sites remains roughly stable after mid-century (RCP 4.5: 191 sites; RCP 8.5: 198), but exposure intensity rises sharply. Total exposed area: RCP 4.5: 16,638 km² (13,192–22,596), ~9.5× baseline; RCP 8.5: 20,969 km² (15,168–28,051), ~12× baseline. Average fraction of site area exposed: RCP 4.5: 11.2% (9.4–14.2); RCP 8.5: 13.7% (10.7–18.3). Very highly exposed sites (>75% exposed) increase to 15 (14–20) under RCP 4.5 and 20 (17–30) under RCP 8.5, a five- to sixfold rise from present.
• Natural vs cultural: By 2100, exposed natural areas dominate due to larger total area: RCP 4.5 natural 15,053 km² (11,906–20,545) vs cultural 1,585 km² (1,287–2,051); RCP 8.5 natural 18,930 km² (13,652–25,545) vs cultural 2,039 km² (1,516–2,507). Mean exposure under RCP 8.5: natural 15.0% (11.6–20.4) vs cultural 9.7% (7.7–12.0).
• Mitigation benefits: Moving from RCP 8.5 to RCP 4.5 reduces median total exposed area by ~21% in 2100 and results in 25% fewer very highly exposed sites by century’s end.
• Country highlights (median projections): By 2100 under both scenarios, all coastal heritage sites are exposed in Cameroon, Republic of the Congo, Djibouti, Western Sahara, Libya, Mozambique, Mauritania and Namibia. Under RCP 8.5 (95th percentile), Côte d’Ivoire, Cabo Verde, Sudan and Tanzania also reach full exposure. Morocco and Tunisia have the highest numbers of exposed sites by 2100 (≥20 each). Countries with largest exposed areas include Mozambique (>5,683 km² under RCP 4.5), Senegal (>2,291 km²), Mauritania (>1,764 km²) and Kenya (>822 km²). Tanzania, Mozambique, Côte d’Ivoire, Benin, Togo and South Africa are projected to have at least 100× more exposed heritage area than at present. Under RCP 8.5, by 2100 Ghana, Sierra Leone, Libya, Mozambique and Seychelles have ~51%, 30%, 25%, 21% and 20% of heritage area exposed, respectively.
• Site-specific examples: SIDS are particularly at risk. Curral Velho (Cabo Verde) reaches 44% exposure (RCP 8.5, 2100). Aldabra Atoll (Seychelles) up to 17%; Kunta Kinteh Island (The Gambia) up to 46% (RCP 8.5, 2100). Cultural/archaeological sites with notable projected exposure (RCP 8.5, 2100): North Sinai Archaeological Sites Zone (Egypt) 91%; Agglomération Aného-Glidji (Togo) 37%; Tipasa (Algeria) 11%; Sabratha (Libya) 7.7%; Carthage (Tunisia) 5.9%.
• Hazard contributions: Flooding contributes more exposed area than erosion overall, though relative importance is site-specific. Natural systems may recover from episodic flooding but can be severely impacted by erosion or changes in salinity and sediment supply; degradation of coral reefs, seagrass and mangroves may weaken natural protection and amplify risks.

The study directly addresses the question of how many and which African coastal heritage sites are exposed to 100-year coastal flooding and erosion now and under future climate scenarios. Findings reveal that one in five coastal AHS are already exposed, with the number of exposed sites tripling by mid-century and exposure severity accelerating thereafter, particularly under high emissions. This underscores the urgency of adaptation for both cultural and natural heritage. Differences in vulnerability between cultural (often built and archaeological) and natural/bio-cultural sites imply tailored risk management: cultural sites can be damaged by both inundation and erosion, while some natural systems can adjust to episodic flooding but may be compromised by chronic erosion and ecosystem degradation. The work highlights co-benefits of emissions mitigation, with substantial reductions in highly exposed sites and total exposed area under RCP 4.5 compared to RCP 8.5. The results provide actionable insights to prioritize countries and sites for protection and to inform governance, monitoring, and site-specific adaptation planning, including hybrid nature-based and engineered solutions. They also highlight cascading ecological and social implications where protective ecosystems (coral reefs, mangroves, seagrasses) are themselves threatened.

This paper delivers the first continent-wide, digitized geospatial assessment of African coastal heritage exposure to combined coastal flooding and erosion under climate change. It quantifies current exposure and projects substantial increases in exposure severity through the century, particularly under high emissions, while demonstrating significant benefits of mitigation. The analysis identifies priority countries and sites, emphasizing the need for immediate adaptation measures, improved governance and monitoring, and consideration of hybrid protective strategies that include ecological infrastructure. Future research should broaden risk assessments beyond coasts to include inland heritage, incorporate additional and compound climate hazards, and evaluate risks arising from adaptation responses (e.g., managed retreat, migration, ecosystem-based adaptation), as well as integrate socio-cultural and economic valuation of heritage, including intangible dimensions.

• Scope restriction: Focuses on coastal hazards (flooding and erosion) only; does not assess other hazards (e.g., riverine floods, heatwaves, wildfires, earthquakes) or compound events with terrestrial/groundwater processes.
• Indicator choice: Uses exposed area as the primary indicator; lacks data to quantify economic losses or intangible cultural values and intra-site heterogeneity of valued components.
• Data and model uncertainty: Digital elevation models (GLO-90) and large-scale inundation modelling introduce uncertainties; coastal morphological projections (erosion/retreat) remain challenging and sensitive to human interventions and sediment budgets.
• Protection assumptions: Assumes no site has 100-year coastal flood protection; site-specific defences may alter local outcomes.
• Hazard components: Marine flooding model omits river discharge and precipitation effects, as well as groundwater processes (subsidence/salinity) that could influence exposure.
• Continental, first-pass assessment: Results are suitable for prioritization and comparison but not as precise local-scale estimates; detailed local studies are needed for site-specific planning.